First, download models for segmentation and hashing
cd models
./get_models.sh
cd ..
Build Docker image to run ONNX models
cd docker
git clone https://github.com/microsoft/onnxruntime.git
docker build -t rbss/onnx:1.0 .
cd ..
Run image with jupyter notebook locally (NVIDIA GPU required)
docker run -v $(pwd):/data -p 8888:8888 --runtime=nvidia -it --rm rbss/onnx:1.0
Got to http://localhost:8888/lab